The Discovery Channel: Getting Content Discovered

Page 1 of 3

Think you are giving customers what they want? Not if they have to navigate through multiple menus and sift through search results to find it. And what if the customer isn't entirely sure what she needs in the first place? If companies want to connect users with content, then they need to remove the pain from the discovery process and provide users with what they want—perhaps even before they know they need it.

Content promises to do everything from entertain to inform and, of course, help grow revenues, yet it poses a challenge that may bring companies to their knees. Offering content seems simple, but as more and more of it becomes available, delivering it effectively becomes a difficult task. Recent research from Nielsen Norman Group highlights the scale of the problem as well as the gross mismatch between user expectations and reality. Search has become the most important site interface element, with more than half of all users going straight for the search button when they hit a Web site. However, these search-dominant users are often sorely disappointed when they fail to get relevant results or worse, get hopelessly lost.

This disconnect occurs because search places the burden of finding relevant information squarely on the shoulders of users, who have to know where to look and be familiar enough with the site's offerings to know what to look for. But search-savvy users are in the minority, according to Nielsen Norman findings. After studying many groups, it concluded that most users are "very poor at query reformulation" and recommended this skill set be part of school curriculum to close the skills gap going forward.

Indeed, expecting users to get to the information they need, when they may not know how to search for it in the first place, is like asking them to look in the dictionary for words they cannot spell (or worse, don't even know). No wonder more than half of all searches are abandoned. Granted, search engines can deliver results from the vast unstructured Internet that are in some way related to the query term, but this method offers neither predictability nor responsibility for quality. As a result, users waste time (and corporate cash) not finding what they're looking for . . . and may never discover what they don't already know.

The Personal Touch
It would be better if relevant content could somehow find users who need it, instead of the other way around. Granted, RSS feeds and news reading technologies that deliver information directly to subscribers as it is published come close. But even with RSS, the burden is on the user to configure the feeds and know the relevant channels to read in the first place.

Findory is a Seattle-based start-up company determined to turn that model on its head by pairing personalization and search. Its patent-pending technology personalizes the site's homepage for each reader, recommending content based on what they've read and what new content is being published. "Anybody can promote a single news agency's content, or build an overly-complicated RSS aggregator. That's not the hard part," observes Greg Linden, Findory's founder. "The hard part is finding interesting and relevant content. We want to provide the best news reading experience based on innovative personalization technology."

Because it is completely user-sensitive, Findory effectively adapts users' homepages as they click around the site. It "learns" from the news users read, searches thousands of feeds and sources, and builds a personalized front page. "We crawl through thousands of news sites, blogs, and articles so users don't have to. All they have to do is read; we do the rest," Linden says. "It doesn't replace search, it does one better because with personalization, users can discover relevant information they didn't know about and didn't know they were looking for."

Predictably, Linden—who wrote the first recommendation engine used by, and later led the software team that developed's personalization systems—has added collaborative filtering to Findory's capabilities mix. "The algorithms do the work, but really it comes down to people," Linden says. "By monitoring what people like you are reading and finding, we can make the match and share that information between users" with similar interests.

Beyond news and enabling relevant content discovery, Findory has also developed an advertising engine that monitors what content users access at Findory and then feeds them relevant ads. For example, if a user reads an article about Harley-Davidson, then the next time the user visits Findory, she will be presented with sponsored links and targeted ads about motorcycle products. "We're basically trying to get to the point where advertising is so relevant that it actually becomes useful," Linden says.

Your Content is Watching
To help simplify the dizzying world of information resources, veteran content provider Ovid Technologies recently developed Ovid SearchSolver, a next-generation federated search solution. This resource discovery tool allows users to conduct a single search across a variety of content sources for uncovering the full breadth of institutionally available content resources.

"It's not about search that delivers thousands of results; it's about precision search that gives back the magic number of most relevant results," observes Diana Bittern, director of technology product management for Ovid. To this end, Ovid is currently expanding on its precision search capabilities to provide even novice searchers with relevant results—and a few surprises.

The next generation of the Ovid platform will be able to "sense" the intent of the user and respond with the information the user needs, as well as information she might not have known she was looking for, Bittern explains. Put simply, this platform will include tools that follow the user, monitor what a user searches for and how, and learns from these patterns to deliver the most relevant results. "It's about providing users a full experience from beginning to end, from discovery to precision search to full text. This completes the food chain of the e resource management cycle, allowing users to discover content and link out to full text wherever it resides."

The company is also eying opportunities to offer recommendation capabilities to link users according to their search habits and information preferences. "You create knowledge by sharing information, and we believe you can also create knowledge by sharing what users search for," Bittern explains. However, adding this capability would require a mechanism to ensure users have control over which searches, if any, are made publicly available.

Page 1 of 3